Performance and efficiency

I understand that some operations are not to be made on high poly-object, included rigging and uv mapping. But the artist may want in first intent to build many details, especially if the model is to be machined full size. The process of rigging can be done later, after the topology has been corrected, and the model transformed, using multi-res and shrink wrap for example. That is different step that can be done by different persons.

1 - Many operations in Blender just start as soon as it is requested, with the parameters selected as in a previous operation, or by default. For example, in Edit mode of a Mesh, Select by all trait/ Non manifold, fires as soon as you selected it. By default all types are selected (Extend, Wires, Boundaries…) With high poly meshes, it may take minutes before the job is done. Then you can deselect what you do not need, and every time you deselect one thing, the operation is triggered again, and again until you get only what you desire. Same with Face intersect, you do not have the opportunity to select the options before the operation is done. Many other operations behave the same, most in fact. It would be great to have a validation button in order to select exactly what you want before the operation is done.

It is not a problem with “small meshes” but it is a real pain with large mesh.

2 - The same happens with modifiers, as boolean, if, before selection the operand object, you forget to select union or intersect that is the operation you want to do, or forget to select some other options, then the modifier is working for nothing. An option “validate” set to false when the modifier is created would let the possibility to the user to set all the parameters he wants. When the validation parameter is activated then Blender can trigger it whenever required as it does now.

3 - The same occurs also in node editor. While modifying connecting changing parameters it may end in very long processing.

4 - I noticed is that not all operation use the power of the computer. For example the plug in “3D Print” works one one core, when in my case there are 31 others available. That is not the only case. I understand that it may be difficult to organize parallel processing I, some cases, but whenever it is possible the use of the full power of the computer is preferable.

5 - When Blender is working, there is no mean to know what it is doing, but remembering what you did last, and there is no way to abort the operation. For example associate armature with automatic weight, with to many vertices may crash Blender. If you select this menu by accident, you just have to wait the end sometimes some hours later, , or wait Blender to crash, or kill Blender, hoping you saved your work. Again that is an example. It would be great to have somewhere the information on what blender is doing, maybe in a modal window, that would also provide a progress information, and a button to abort. A framework should be implemented, and, the different potentially long operations could in the future use this framework.

This could be used for modifiers when set up by the user the first time, and for single shot operations as select by all trait, 3D Print operations, rigging.